摘要：CuBi2O4作為最有前景的光陰極材料之一，其理論光電流密度可達(dá)20 mA·cm?2。然而，在實(shí)際光電化學(xué)反應(yīng)中，由于光生載流子復(fù)合嚴(yán)重，實(shí)際的光電流密度相較于理論值仍然存在顯著差異。一般而言，光電化學(xué)性能在很大程度上依賴于光生載流子的高效分離和傳輸，以及快速的反應(yīng)動(dòng)力學(xué)。本文中，我們提出了一種多金屬氧酸鹽（多酸）修飾的CuBi2O4/Mg-CuBi2O4同質(zhì)結(jié)光陰極。系統(tǒng)考慮了光陰極體相和界面的載流子傳輸：首先，通過(guò)CuBi2O4/Mg-CuBi2O4同質(zhì)結(jié)中所構(gòu)造的內(nèi)建電場(chǎng)實(shí)現(xiàn)光生電子和空穴的定向轉(zhuǎn)移；另外，多酸助催化劑Ag6[P2W18O62]（AgP2W18）在反應(yīng)過(guò)程中可被還原，進(jìn)而可被用作質(zhì)子存儲(chǔ)載體，可在抑制載流子復(fù)合的同時(shí)促進(jìn)界面光電化學(xué)反應(yīng)。這種協(xié)同作用可在促進(jìn)體相-界面載流子傳輸?shù)耐瑫r(shí)解決界面緩慢反應(yīng)動(dòng)力學(xué)這一瓶頸。結(jié)果表明，本文所提出的光陰極實(shí)現(xiàn)了出色的光電化學(xué)性能，在0.3 V vs. RHE時(shí)，光電流密度達(dá)?0.64 mA·cm?2；而在使用H2O2電子犧牲劑后，相同電位下的光電流密度進(jìn)一步提升到?3 mA·cm?2。本文所提出的光陰極與已有研究工作中報(bào)道的最佳結(jié)果相比，具有相當(dāng)?shù)墓怆娀瘜W(xué)活性，證明了其在實(shí)際應(yīng)用中的巨大潛力。

關(guān)鍵詞：鉍酸銅；鎂摻雜；光陰極；多酸助催化劑；光電化學(xué)

中圖分類號(hào)：O644

Abstract： Photoelectrochemical water splitting using semiconductormaterials is one of the most promising methods for converting solarenergy into chemical energy. Among the commonly usedsemiconductors， p-type CuBi2O4 is considered one of the most suitablephotocathode materials and can allow a theoretical photocurrent densityof about 20 mA?cm?2 for photoelectrochemical water splitting. However，due to severe charge carrier recombination， the obtained photocurrentdensity is much lower than the theoretical value. Highly efficientphotoelectrochemical performance relies on fast charge carrierseparation and transport， and prompt reaction kinetics. In this study， wereport the development of a polyoxometalate-modified CuBi2O4/Mg-CuBi2O4 homojunction photocathode to improve both the bulk and interfacial charge carrier transport in the photocathode.For the bulk of the photocathode， the built-in electric field originating from the CuBi2O4/Mg-CuBi2O4 homojunction promotesthe migration of photo-excited electrons on the conduction band from pure CuBi2O4 to Mg-doped CuBi2O4. Additionally，the electric field facilitates the transfer of holes from the valence band of Mg-doped CuBi2O4 to pure CuBi2O4. Thisdirectional transfer of both photo-excited electrons and holes plays a significant role in promoting separation andsuppressing the recombination of the charge carriers. On the surface of the photocathode， the reduced polyoxometalateco-catalyst Ag6[P2W18O62] （AgP2W18） was used as a proton sponge to accelerate surface reaction kinetics and suppresscarrier recombination. These synergistic effects improved the photo-generated charge carrier transfer and reaction kinetics.As a result， the novel photocathode displayed excellent photoelectrochemical properties， and the photocurrent density wasobserved to be ?0.64 mA?cm?2 at 0.3 V vs. RHE， which is better than that of ?0.39 mA?cm?2 for a pure photocathode.Furthermore， the novel photocathode had an applied bias photon-to-current efficiency （ABPE） higher than 0.19% at 0.3 Vvs. RHE. In contrast， the pure photocathode had an ABPE of ～0.12% under the same conditions. Additionally， when H2O2was used as an electron scavenger， the photocurrent density was ?3 mA?cm?2 at 0.3 V vs. RHE， which is an improvementof approximately 1.5 times compared to the pure photocathode. Furthermore， the charge separation and charge injectionefficiency of the novel photocathode were significantly improved compared with the pure photocathode. The experimentalresults conclusively indicate that the formation of the CuBi2O4/Mg-CuBi2O4 homojunction and AgP2W18 modification playeda significant role in the improved performance of the CuBi2O4 photocathode. The performance of the novel photocathodewas comparable with the results reported in previous studies， demonstrating its promising potential in real applications.

Key Words： CuBi2O4; Mg-doping; Photocathode; Polyoxometalate co-catalyst; Photoelectrochemistry

1 Introduction

CuBi2O4 nanomaterial is a promising p-type semiconductorbecause it inherently emerges the following merits of （i） widespectral absorption range （1.4–1.8 eV）， （ii） suitable band edgepotential， and （iii） high starting voltage 1–3. These salient meritsmake CuBi2O4 nanomaterial can utilize most of the visible lightin the solar spectrum and present excellent potentiality inphotoelectrochemical water splitting. Assuming that all photonswith energy higher than the band-gap of CuBi2O4 can beabsorbed and fully utilized， under the illumination of simulatedAM 1.5G spectra， the theoretical maximum photocurrent densitycan reach ～20 mA·cm?2 4. However， the reported results showthat the photocurrent densities of pure CuBi2O4 photocathodesare lower than the theoretical value 5. This is because theCuBi2O4 materials still suffer from some disadvantagesincluding severe surface recombination， poor charge separationand transport， slow reaction kinetics， etc.， both of these defectsthen significantly inhibit the improvement in solar energyconversion and utilization efficiency 6–8.

To solve these issues， various strategies have been proposedin the past decades 9. A feasible solution is to construct a builtinelectric field to improve the charge carrier separation inCuBi2O4 10. Here， the built-in electric field is usually achievedthrough a junction that is formed by direct contact between thetwo semiconductors with different Fermi levels. The existenceof junction results in band bending in the depletion region withcharge carrier redistribution. For example， the preparation ofNiO/CuBi2O4 or CuBi2O4/CuO heterojunction photocathodeshows better performance and stability. With this heterojunction，photo-generated holes and electrons can efficiently transfer tothe NiO layer and the CuO layer， respectively， resulting in fasterseparation and migration of charge carriers， and thus improvedthe photoelectrochemical performance 11，12. Meanwhile， usingtwo identical semiconductor layers， where one layer wassubjected to ion doping to change the Fermi level， to construct ahomojunction is also an alternative method 13，14. A typical casewas proposed by Zhang et al. 15， where the gradient of the bandedge energetics photoelectrode was fabricated by using Mg2+ asthe dopant. Firstly， for some p-type semiconductors likeCuBi2O4， the Mg2+ doping can indeed shift the Fermi level 16，17.In addition， since the oxidation state of Mg2+ is lower than Bi3+in CuBi2O4， the carrier concentration may be increased afterdoping 18. It should also be pointed out that the formation of ajunction is a strategy for enhancing charge carrier transfer in theCuBi2O4 bulk phase， whereas the photoelectrochemical reactiononly occurs at the CuBi2O4-electrolyte interface. The possibletrapping sites on the surface and slow reaction kinetics then willrestrain the interfacial charge carrier transfer and utilization， thusimpairing the photoelectrochemical performance 19. To improvethe interfacial charge carrier transfer and facilitate the reaction，surface modification using co-catalyst has been demonstrated asa practicable strategy 20. To date， a variety of co-catalysts havebeen proposed， including noble and non-noble metals， enzymes，polyoxometalates， etc. 21，22. Among them， polyoxometalatebasedco-catalyst is regarded as a promising choice because itinherently exhibits excellent electron transfer ability， superiorsurface reaction kinetics， etc. 23. As a result， polyoxometalateshave been used in various scenarios， such as photocatalytichydrogen production 24，25， degradation of organic pollutants 26，27，CO2 reduction 28，29 and N2 reduction 30，31.

Herein， inspired by the homojunction design along with themerits offered by the polyoxometalate-based co-catalyst， a polyoxometalate modified CuBi2O4/Mg-CuBi2O4 homojunctionphotocathode was proposed for photoelectrochemical hydrogenevolution reaction. In detail， the homojunction between pureCuBi2O4 and Mg-doped CuBi2O4 was first fabricated on FTOglass， and then the polyoxometalate-based co-catalyst ofAg6[P2W18O62] （AgP2W18） was deposited on the surface ofCuBi2O4/Mg-CuBi2O4 homojunction. As shown in Scheme 1，the built-in electric field originated from the CuBi2O4/Mg-CuBi2O4 homojunction not only promotes the migration ofphoto-excited electrons on conduction band from pure CuBi2O4to Mg-doped CuBi2O4 but also facilitates the transfer of holes onvalence band from Mg-doped CuBi2O4 to pure CuBi2O4.Besides， since AgP2W18 can be reduced during the reaction， thereduced [P2W18O62]6? then can store protons to satisfy therequirement of electrical neutrality 32，33. The storage of protonswill play a positive role in promoting the interfacialphotoelectrochemical reaction and thus the efficient utilizationof the surface charge carriers 34，35. Such synergetic effectbetween the bulk and surface transport then suppressed therecombination of photo-generated charge carriers and improvedsurface activity. By leveraging this strategy， the excellentphotoelectrochemical performance of the proposedphotocathode was achieved， where the photocurrent densityreached ?3 mA·cm?2 at 0.3 V vs. RHE with H2O2 electronscavenger. This performance is comparable to the best resultsreported in previous works， further indicating the noticeablenatural endowment of the proposed photocathode structure inefficient solar energy conversion and utilization.

2 Experimental section

2.1 Materials

All chemicals were used as received and without furtherpurification. Bismuth nitrate pentahydrate （Bi（NO3）3·5H2O，99.99%）， copper nitrate trihydrate （Cu（NO3）2·3H2O， 99.99%），silver nitrate （AgNO3， analytical grade）， polyethylene glycol（average Mn 20000， PEG-20000， analytical grade）， absoluteethanol， ethylene glycol， and acetic acid were purchased from Aladdin. These three liquid reagents are analytical grade.Magnesium nitrate （Mg（NO3）2·2H2O， analytical grade） waspurchased from Sinopharm Chemical Reagent Co.， Ltd.Fluorine-doped tin oxide （FTO） glass （7 ·square?1） was fromWuhan Jinge Solar Energy Technology Co.， Ltd. TheK6[P2W18O62] was synthesized according to a previous method，and the Infrared and Raman spectra results shown in Figs. S1and S2 demonstrated that K6[P2W18O62] was successfullysynthesized 36，37. Deionized water used in the whole experimentwas obtained from a ULUPURE UPT-11-20T Ultrapure WaterSystem for the laboratory.

2.2 Fabrication of the Mg-doped CuBi2O4/AgP2W18photocathode

Fig. 1 schematically illustrated the fabrication procedure ofthe Mg-doped CuBi2O4/AgP2W18 photocathode. Here， thepreparation of the proposed photocathode can be divided into thefollowing three steps. Firstly， the CuBi2O4 film was fabricatedon FTO through a modified spin-coating method 38. Where，Bi（NO3）3·5H2O （1.08 mmol）， Cu（NO3）2·3H2O （0.54 mmol）， andPEG-20000 （200–500 mg） were dissolved in a mixture solution，which contained 1 mL absolute ethanol， 1 mL ethylene glycol，and 1 mL acetic acid， to obtain a clear and colorless pureCuBi2O4 precursor solution. The pure precursor solution wasspun onto the FTO substrate with a speed of 1500 r·min?1 for 15s and 2500 r·min?1 for 40 s， and heated at 150 °C for 10 min.Then， （Mg（NO3）2·2H2O was dissolved into the pure CuBi2O4precursor to prepare the doping precursor solution. Here， themolar concentration of Mg（NO3）2 was 5% of that of CuBi2O4 inthe precursor solution. Using the same spin-coating method asabove， the Mg-doped CuBi2O4 film was prepared on the top ofpure CuBi2O4 film. After annealing at 550 °C for 120 min in anatmosphere furnace， a homojunction between the pure and dopedCuBi2O4 films was obtained. In this work， the correspondingphotocathode was denoted as Mg-CuBi2O4. Finally， quantitativeK6[P2W18O62] and corresponding AgNO3 were dissolved indeionized water to obtain Ag6[P2W18O62] nanoparticles （namedas AgP2W18） colloid， which could be recognized by scanningelectron microscopy （SEM） image and Tyndall effect （see Fig.S3）. The AgP2W18 was spin-coated on the top of the Mg-CuBi2O4 （2000 r·min?1 for 30 s） and the composite film wasnamed as Mg-CuBi2O4/AgP2W18 photocathode. In addition， thepure CuBi2O4 photocathode was also fabricated through the same spin-coating method. But for a fair comparison， the firststep， i.e. the preparation of pure CuBi2O4 film on FTO glass， wasrepeated once.

2.3 Characterization

X-ray diffraction （XRD， Bruker-AXS D8 Advance， Germany）analyses were recorded with Cu-K radiation （ = 0.15418nm， 40 kV， 30 mA）. Scanning electron microscopy （SEM， Zeiss，EVO 10， Germany） was used to observe the morphology. Theenergy dispersive spectrometer （EDS， Oxford Xplore， UK） wasattached to the SEM. Transmission electron microscopy （TEM，F(xiàn)EI Tecnai G2 F20， USA） was used to investigate themicrostructure details of the samples. The X-ray photoelectronspectra （XPS） measurements were performed by a spectrometer（Thermo Scientific ESCALAB Xi+， USA） with Al-K radiationsource （hν = 1486.6 eV）. The UV-Vis absorbance spectroscopywas recorded with a UV-Vis Spectrophotometer （Shimadzu UV-3600， Japan）. Infrared spectrum （IR， Nicolet Magna 560 FT-IR，Agilent， USA） was employed in this work. Raman spectroscopy（532 nm， HR-Evolution， Horiba， France） was used to analyze thechemical structure of the proposed photocathode. Transientsurface photovoltage （TSPV， Perfectlight， China） measurementswere carried out under a 355 nm laser pulse.

2.4 Photoelectrochemical measurements

The photoelectrochemical measurements were performed byan electrochemical workstation （CHI660E， Shanghai ChenhuaInstrument Corp.）. Schematic illustration of thephotoelectrochemical testing system was shown in Fig. S4. A300 W Xe lamp （AM 1.5G filter） was used to simulate thesunlight illuminationwith the light intensity of 100 mW·cm?2.Then， 0.3 mol·L?1 K2SO4 and 0.2 mol·L?1 phosphate buffer （pH6.5） was used as the electrolyte. The photocathodes wereilluminated from the film side unless otherwise mentioned. Ptsheet was employed as the counter electrode and Ag/AgClelectrode （saturated KCl） was selected as the reference electrode.The measured potentials vs. Ag/AgCl were converted against areversible hydrogen electrode （RHE） scale with Eq. （1）：

RHE E ＝ EAg/AgCl ＋ 0.0592pH ＋ 0.197 （V） （1）

where EAg/AgCl is the obtained potential. The current-potentialplots were measured in a voltage window ranging from 0.3 to1.2 V vs. RHE a scanning rate of 50 mV·s?1. The current-timeplots were carried out at a constant bias of 0.4 V vs. RHE.Electrochemical impedance spectroscopy （EIS） was performedat 0.4 V vs. RHE under illumination （5 mV AC voltageamplitude， frequency range： 0.1 Hz–100 kHz）. The Mott-Schottky plot was obtained in in the dark conditions （increment：5 mV， frequency： 1 kHz）.

3 Results and discussion

3.1 Characterization of Mg-doped CuBi2O4/AgP2W18 photocathode

Here， considering that the amount of the template agent PEG-20000 added in the pure CuBi2O4 precursor solution will have asignificant impact on the morphology of the CuBi2O4 film 39， we have first characterized the corresponding surface structures andphotoelectrochemical performances of the original CuBi2O4films with different PEG-20000 contents. The results shown inFig. S5 indicated that a relatively flat surface with a few smallpores was obtained at low PEG-20000 content of 200 mg，whereas the evident nanoporous structure with a large number oftunnels can be obtained with the increase of PEG-20000 content（Figs. 2a and S5）. Since the nanoporous structure of CuBi2O4film can offer more active site to present a positive role inpromoting the separation of photo-generated carriers， a betterphotoelectrochemical performance then can be obtained with theincrease of PEG-20000 dosage （Fig. S5）. However， it shouldalso be pointed out that excessive addition of PEG-20000 canmake the precursor solution too viscous to spread and form a thinfilm on FTO glass， thus in the following experiments， the contentof PEG-20000 was all set at 500 mg. Fig. 2b，c showed the SEMimages of the Mg-doped CuBi2O4 film and Mg-dopedCuBi2O4/AgP2W18 film. As can be seen， both of these twosamples can maintain an obvious nanoporous structure，indicating that Mg-doping and AgP2W18 loading did not alter themorphology of the CuBi2O4 film.

The precise structure of the above three samples was alsoinvestigated by HRTEM and the results were shown in Figs. 2e，S6 and S7. The HRTEM image showed that the lattice spacingof Mg-CuBi2O4/AgP2W18 was 0.268 nm， which corresponded tothe （130） and tetragonal phase CuBi2O4 （Fig. 2e）. Besides，AgP2W18 nanoparticles adjacent to the CuBi2O4 were alsoobserved， indicated that the CuBi2O4 and AgP2W18 werecombined successfully. It can also be found that there was noobvious lattice of nanoparticles in Fig. 2e， which means that theAgP2W18 nanoparticles are amorphous. Meanwhile， theelemental distributions of： i） Mg except Bi， Cu， and O inelemental mapping profiles in Figs. S6 and S7， and ii） Ag， P， andW except Mg， Bi， Cu， and O in Figs. 2f and S7 also implied theformation of Mg-doped CuBi2O4 film and AgP2W18-loaded Mg-CuBi2O4 film， respectively. In addition， the Mg 1s signals wereonly detected on the Mg-CuBi2O4 film and Mg-CuBi2O4/AgP2W18 film （Fig. S8）， which could further confirmthe successful doping of the Mg element.

Fig. 3a shows the XRD pattern of the original CuBi2O4 film，various diffraction peaks at 20.9°， 28.1°， 33.4°， and 46.8°comparable to those of the tetragonal phase CuBi2O4 （PDF#71-1774） （200）， （211）， （130） and （141） crystal planes were welldistinguished 3，7. Here， we also calculated the crystallite size ofCuBi2O4 according to the Scherrer Eq. （2） 40，41：

where， L is the particle size and K is a dimensionless shapefactor. λ is the X-ray wavelength 0.15418 nm， β is the linebroadening at half the maximum intensity （FWHM）， and θ is theBragg angle. The calculated results showed that the averagecrystallite size of CuBi2O4 is about 95 nm， which is consistentwith the HRTEM results. Compared with the pure CuBi2O4， it can be seen that the introduction of Mg led to a slight angularshift in the （211） plane， indicating an effective doping of Mg inthe CuBi2O4 lattice 42. Whereas for the AgP2W18 loading， asimilar shift in the （211） plane was also observed， whichrevealed that the loading of AgP2W18 has not affected thecrystalline structure of the Mg-doped CuBi2O4 sample.Moreover， we have also measured the chemical compositions ofthe above three samples using the Raman spectrum. Resultsshown in Fig. 3b indicated that there existed four prominentpeaks centered at 130.8， 262.3， 404.2， and 586.1 cm?1 for pure CuBi2O4. These four peaks match with the A1g mode and arerelated to the CuBi2O4 tetragonal structure. Where， those peakscorresponded to translational vibrations of the CuO4 along the zaxis（130.8 cm?1）， rotation of two stacked CuO4 squares inopposite directions （262.3 cm?1）， stretching vibration mode ofthe Bi―O bond （404.2 cm?1） and breathing vibration mode inthe CuO4 square （586.1 cm?1）， respectively. But after Mgdoping， a slight shift of CuO4 peak was also observed （Fig. 3b）.This peak shift could be ascribed to the lattice distortion 43，44，which also indicated the successful doping of Mg into CuBi2O4films 6，45. When AgP2W18 was further loaded on the surface ofMg-CuBi2O4， no significant change was recognized， revealingthat the co-catalyst loading will not influence the chemicalstructure of Mg-doped CuBi2O4. These results were alsoconsistent with the XRD results.

3.2 Performance evaluation of Mg-doped CuBi2O4/AgP2W18 photocathode

In this section， the photoelectrochemical hydrogen productionperformance of both pure CuBi2O4 and Mg-doped CuBi2O4 aswell as AgP2W18 nanoparticle loaded Mg-CuBi2O4photocathodes were evaluated in a three-electrode system. Fig.4a showed the comparison of the current-voltage curves， whichwere established by linear sweep voltammetry， of the abovethree photocathodes without the addition of H2O2. Resultsshowed that the photocurrent density of the Mg-CuBi2O4/AgP2W18 photocathode was higher than that of theoriginal CuBi2O4 photocathode and Mg-CuBi2O4 photocathodein the voltage window from 1.2 to 0.3 V vs. RHE. Where， thephotocurrent density of the proposed photocathode was about?0.64 mA?cm?2 at 0.3 V vs. RHE， which was better than that of?0.39 mA?cm?2 for pure CuBi2O4 photocathode and ?0.51mA?cm?2 for Mg-CuBi2O4 photocathode. Here， considering the co-catalyst AgP2W18 also presented photoresponse， to moreclearly clarify its role， we have also tested the linear sweepvoltammetry （LSV） curves of pure AgP2W18 （Fig. S9）. Thephotocurrent of AgP2W18 is very small compared with that of theCuBi2O4 photocathodes. This intensification may be attributedto the following two reasons. On one hand， after Mg doping， thefabricated CuBi2O4/Mg-CuBi2O4 homojunction also constructeda built-in electric field to improve charge carrier transfer andsuppress the recombination. On the other hand， sinceAg6[P2W18O62] can be reduced during the reaction， to satisfy therequirement of electrical neutrality， additional protons will beenriched in the co-catalyst. This enrichment indicated that thereduced Ag6[P2W18O62] can be used as a proton sponge tocontribute to the proton reduction to hydrogen. These two meritsnot only promoted the charge carrier separation and transport butalso relieved the sluggish reaction kinetics， thus leading to theoptimum photoelectrochemical performance of the proposedMg-CuBi2O4/AgP2W18 photocathode. Applied bias photon-tocurrentefficiency （ABPE） can be calculated according to theEq. （3）： 46

where jph is the photocurrent density， Eapplied is the applied bias，and Ptotal is the light density of 100 mW?cm?2. The maximum ofABPE for pure CuBi2O4 was ～0.12% at 0.3 V vs. RHE， and Mg-CuBi2O4 exhibited a higher ABPE of 0.15% at 0.3 V vs. RHE.Mg-CuBi2O4/AgP2W18 exhibited a highest value of 0.19% at 0.3V vs. RHE （Fig. S10a）. It can be assumed that the CuBi2O4/Mg-CuBi2O4 homojunction and AgP2W18 can contribute to theelectron transport.

Charge carriers recombination behavior was investigated byphotocurrent density-time curves during the light-on-offprocesses （Fig. 4b）. It was obvious that the three photocathodesexhibited transient photocurrent spikes due to the severerecombination of photogenerated carriers that accumulatedunder illumination. Compared with the pure CuBi2O4photocathode， the Mg-CuBi2O4 photocathode showedsignificant enhancement in photocurrent density due to itsprompt transfer of accumulated holes and suppression of severeback recombination. Among the three photocathodes， the Mg-CuBi2O4/AgP2W18 photocathode revealed the highestphotocurrent density. These results were consistent with thephotoelectrochemical activity observed in Fig. 4a. Moreover， thetransient decay time was obtained by the Eq. （4）： 47

where the time at lnD = ?1 was assumed to be the transient timeconstant， It， Is， and Ip are the current at time t， the steady-stateand spike current， respectively. In Fig. 4c， after Mg doping， theCuBi2O4/Mg-CuBi2O4 homojunction demonstrated a longertransient decay time for the Mg-CuBi2O4 photocathodecompared to the pure CuBi2O4 photocathode. This suggests adecrease in the charge recombination rate. When AgP2W18 wasdeposited on the surface of Mg-CuBi2O4 photocathode， a largertransient decay time indicated the suppressed chargerecombination. This effect might be attributed to the rapid reaction kinetics of AgP2W18. All of these three photocathodespresented a decay photocurrent， but the proposed photocathodesimultaneously emerged with the highest photocurrent densityand relatively excellent stability. Meanwhile， we alsocharacterized the physical and chemical properties of both freshand tested samples. SEM images revealed that， after thephotoelectrochemical reaction， additional micron-scalecomponents were generated on the electrode surface， but thenanoporous structure was still retained for each photocathode（Fig. S11）. XRD and Raman analysis indicated that thecrystalline structures after the reaction of both pure and Mgdopedas well as AgP2W18 doped photocathodes is similar tothose of fresh ones （Figs. S12 and S13）.

Fig. S14a showed the open circuit potential （OCP）-timecurves of the three photocathodes. The OCP of the threephotocathodes shifted positively under light， which ischaracteristic of a p-type semiconductor and consistent with theMott-Schottky results 48. The larger OCP （difference of opencircuit potential between under light and dark） further confirmedthe enhanced separation of photogenerated charge carriers.Furthermore， the charge carrier recombination lifetime （τ） at themoment of light on/off is quantified by the Eq. （5）： 49

where k is the Boltzmann constant （1.38 × 10?23 J?K?1）， T is thetemperature， and e is the electron charge （1.60 × 10?19 C）. Asshown in Fig. S14b， the calculated τ for pure CuBi2O4photocathode is ca. 148 ms at the transient of light-off. Underthe same conditions， the τ of Mg-CuBi2O4 photocathode is ca.110 ms. The Mg-CuBi2O4/AgP2W18 photocathode exhibitedminimal charge carrier recombination lifetime （ca. 70 ms）. TheMg-CuBi2O4/AgP2W18 photocathode showed faster ΔOCPdecay kinetics， which demonstrated the smaller recombinationfactors. Transient surface photovoltage （TSPV） was used tostudy the dynamics of photogenerated charge carriers for threephotocathodes. Under the same conditions， the photovoltageintensity was directly related to the concentration ofphotogenerated carriers on the surface of the semiconductormaterial 50. Fig. S14c showed the TSPV curves of the threephotocathodes. As shown， after Mg doping， the CuBi2O4/Mg-CuBi2O4 homojunction presented a more negative TSPV signal，which is favorable for photogenerated electrons to gather onsurface of Mg-CuBi2O4 photocathode. Meanwhile， the TSPVsignal intensity of Mg-CuBi2O4/AgP2W18 is higher than that ofMg-CuBi2O4， suggesting the photogenerated electrons can beeffectively extracted into AgP2W18 51. Both OCP and TSPV testsrevealed that the separation efficiency of photogenerated carrierscan be promoted because of the existence of CuBi2O4/Mg-CuBi2O4 homojunction and modification of AgP2W18.

Besides， to more accurately illustrate the potential of theproposed photocathode in photoelectrochemical water splittingto hydrogen， the photoelectrochemical activities of the abovethree photocathodes were also characterized in the presence of electron scavenger H2O2. The obtained results shown in Fig. 4drevealed that the photocurrent densities of each photocathodewere significantly enhanced， while the proposed Mg-CuBi2O4/AgP2W18 photocathode still emerged with the superiorperformance. Where， at the working potential of 0.3 V vs. RHE，the photocurrent density of Mg-CuBi2O4/AgP2W18photocathode was ?3 mA?cm?2， which achieved a remarkableimprovement of about 1.5-times and 1.23-times compared withthe pure CuBi2O4 photocathode （?2.05 mA?cm?2） and Mg-dopedCuBi2O4 photocathode （?2.43 mA?cm?2）. Charge separationefficiency （ηsep） of three photocathodes was measured toinvestigate the effect of Mg-doping and AgP2W18 co-catalyst onthe photoelectrochemical performance. And the surfacebehaviors of photocathode was evaluated by charge injectionefficiency （ηinj）. ηsep， ηinj， and Jabs are calculated using the Eqs.（6）–（9） 52，53：

where Jabs is the photocurrent density converted from theabsorbed photons， JH2O2 and JH2O are the photocurrent densitiesobtained in electrolyte with and without electron scavengerH2O2， respectively. ΦA(chǔ)M1.5（λ） is the photo flux， A（λ） is theabsorptance. The spectra and photo flux were displayed in Fig.S15. After integrating the electron flux from 300–900 nm， Jabswas determined to be ?17.4 mA?cm?2 for pure CuBi2O4photocathode. The calculated ηsep and ηinj values are shown inFig. S10b，c. Obviously， CuBi2O4/Mg-CuBi2O4 homojunctioncould improve the ηsep value of Mg-CuBi2O4 photocathode in thevoltage region. At 0.3 V vs. RHE， the ηsep values of Mg-CuBi2O4/AgP2W18 photocathode reached the highest values of17.4 %. The ηinj of the Mg-CuBi2O4/AgP2W18 photocathode wasenhanced to 22.6% at 0.3 V vs. RHE in Fig. S10c. The resultsindicated that the homojunction between the pure CuBi2O4 andMg-doped CuBi2O4 significantly promoted the separation ofelectrons and holes， and the AgP2W18 could also enhance theseparation. The long-term operation stability results were shownin Fig. S10d， three decay photocurrent was observed on the threephotocathodes during continuous light-on processes， indicatingthat the Mg-CuBi2O4/AgP2W18 photocathode remains relativelybetter stability compared with the other two photocathodes.

Furthermore， we have also made a brief summary of recentadvances on CuBi2O4-based photocathode forphotoelectrochemical hydrogen generation and compared thephotocurrent density between the photocathode reported in theliterature and in this work （Table S1）. As can be seen， comparedwith the best results reported in previous works， our proposedphotocathode can also emerge comparable photoelectrochemical activities under the operation condition of both with and withoutH2O2. These results further indicated the noticeable naturalendowment of the proposed photocathode structure for efficientconversion of solar energy.

3.3 Analysis of the underlying mechanism

In this section， the underlying mechanism of the proposedphotocathode in enhancing the photoelectrochemicalperformance was discussed. Here， we first discussed thevariation in Fermi Level by the Mott-Schottky plots， see Fig. 5a.To calculate the flat-band potential from the Mott-Schottky plotson the basis of the Eq. （10） 9：

where C is the space charge capacitance of semiconductor，relative permittivity （ε） is 80 for CuBi2O4 54， vacuumpermittivity （ε0） is 8.86 × 10?12 F?m?1， A is the area ofphotocathode， Na is the carrier density， T is the temperature（room temperature）， VFB is the flat-band potential， and V is theapplied potential. After the flat band potential is approximatedas the Fermi level 55，56， it can be found that the original flat-bandpotential of pure CuBi2O4 film was 1.21 V vs. RHE， while itgradually shifted from 1.21 V vs. RHE to 1.12 V vs. RHE afterMg doping. From the slope of the plots， the carrier densitiesincreased after doping 57. The carrier densities are calculatedusing the Eq. （11）：

Then， the carrier densities of pure CuBi2O4 and Mg-CuBi2O4 were calculated as 1.37 × 1019 and 3.04 × 1019 cm?3， respectively.Higher carrier density in Mg-doped CuBi2O4 is favorable forhydrogen production activity 58. Meanwhile， based on theabsorbance spectra （Fig. 5b） of Mg-doped CuBi2O4 film， theband-gap was then estimated according to the Eq. （12） 59：

（αhν） 2＝ A（hν - E ） （12）

Where hν， α， A and Eg are photon energy， K-M function， theproportionality constant， and band-gap， respectively. The energyintercept of a plot of （αhν）2 against （hν） yields Eg was shown inFig. 5c and the corresponding band-gaps are 1.6 and 1.57 eV forpure CuBi2O4 and Mg-CuBi2O4， respectively. The cyclevoltammetry curves of pure CuBi2O4 and Mg-CuBi2O4 shown inFig. S16 were used to investigate the lowest unoccupied（LUMO） and the highest occupied molecular orbital （HOMO）.The LUMO/HOMO levels were calculated as ?4.11 V/?5.71 Vand ?4.08 V/?5.65 V for pure CuBi2O4 and Mg-CuBi2O4，respectively. These values can be converted into ?0.33 V/1.27 Vvs. RHE and ?0.36 V/1.21 V vs. RHE. Commonly， LUMO andHOMO are approximately equivalent to the conduction band（CB） and the valence band （VB） 60. Thus， the CB/VB of pureCuBi2O4 and Mg-CuBi2O4 were ?0.33 V/1.27 V vs. RHE and?0.36 V/1.21 V vs. RHE， respectively. It can be found that， forindividual CuBi2O4 and Mg-doped CuBi2O4 films， beforecontact， they possess different Fermi levels， as shown in Fig. 5d（left）. But once they come into contact to form a homojunction，the Fermi levels will reach equilibrium， so the valence band andconduction band bend at the interface， as illustrated in Fig. 5d（right）. The band bending would play a positive role inpromoting the migration of photo-excited electrons to CuBi2O4 surface and holes to FTO side， resulting in enhancing the chargecarrier separation and transport.

To further demonstrate this point， XPS was also employed tomeasure the chemical shift of the three samples. As shown inFig. 6a， the Bi 4f spectra of pure CuBi2O4 presented twocharacteristic peaks centered at 163.8 eV （Bi 4f5/2） and 158.6 eV（Bi 4f7/2）， respectively， which can be assigned to Bi3+ 61，62.However， both of these two peaks of Mg-CuBi2O4 revealed anegative shift in binding energy compared with pure CuBi2O4.Such negative shift can be ascribed to the acceptance ofelectrons 63，64， where the photo-generated electrons wereefficiently moved from pure CuBi2O4 to Mg-doped CuBi2O4 inaccordance with the band bending mechanism 52，65. Meanwhile，it can also be found that a shift to higher energy in Mg-CuBi2O4/AgP2W18 compared with Mg-CuBi2O4 can beobserved， which indicated the electron transfer from Mg-CuBi2O4 to AgP2W18. This directional transfer of electrons is inalignment with the above analysis of the underlying mechanismin the previous section. Meanwhile， we also analyzed the highresolutionCu 2p spectra of the above three samples. As can beseen in Fig. 6b， two characteristic spin-orbit split spectrarevealed two major peaks of Cu2+ in pure CuBi2O4 centered at933.5 eV （Cu 2p1/2） and 953.3 eV （Cu 2p3/2）. Theoretically， theMg doping will lead to a positive shift in binding energy whilethe electron transfer results in a negative shift 4，66. This trade-offthen emerged a slim shift of binding energy of Cu2+ from 933.5eV （Cu 2p1/2） and 953.3 eV （Cu 2p3/2） of pure CuBi2O4 to 933.7eV （Cu 2p1/2） and 953.5 eV （Cu 2p3/2） of Mg-CuBi2O4. However，once the AgP2W18 was loaded on the Mg-CuBi2O4 surface， thefast electron transfer from Mg-CuBi2O4 to AgP2W18 made thepositive shift of binding energy caused by Mg-doping moreobvious. These results further demonstrated the directionalelectron transfer， which then promoted the charge separation andsuppressed the recombination， thus facilitating thephotoelectrochemical performance. Moreover， the XPS resultsshown in Fig. 6c，d also implied that the original Ag+ and W6+ inMg-CuBi2O4/AgP2W18 photocathode can be reduced to Ag0 andW5+ during the photoelectrochemical reaction. The presence ofW5+ indicated that the anion [P2W18O62]6? in AgP2W18 gotelectrons， while the cation Ag+ in AgP2W18 was partiallyreduced. To satisfy the requirement of electrical neutrality， thenegatively charged reduced-AgP2W18 will attract protons ascounter cations for proton storage 33，67. Such dual function ofstoring protons and acting as active sites then can play asignificant role in overcoming the sluggish reaction kinetics andthus enhancing the interfacial photoelectrochemical reaction（Fig. 4）.

Furthermore， the EIS was also measured to clarify the chargetransfer characteristics. As shown in Fig. 6e， according to theequivalent circuit， we fitted the experimental data， and thedetailed information was listed in Table S2. Here， Rs representsthe series resistance between FTO and the CuBi2O4 film. Rbulkand Cbulk are related to the bulk charge transport resistance andspace-charge layer， respectively. Rct and CPE refer to interfacialcharge transfer resistance and surface states capacitance，respectively 68. As shown， Rs for each photocathode was similarand far below the bulk and interfacial charge transfer resistance，indicating that the effect of series resistance can be ignored 69.However， for both Mg-CuBi2O4 and Mg-CuBi2O4/AgP2W18photocathodes， Rbulk and Rct achieved a remarkable decrease.Especially for the co-catalyst AgP2W18 loaded photocathode， thebulk and interfacial charge transfer resistance can be further declined. These results revealed that the Mg-doping andAgP2W18-loading can enhance charge carriers transfer comparedto the pure CuBi2O4， which then played a positive role infacilitating photoelectrochemical reaction.

4 Conclusions

In summary， we prepared a polyoxometalate AgP2W18modified CuBi2O4/Mg-CuBi2O4 homojunction photocathode forhighly efficient photoelectrochemical hydrogen evolutionreaction. Ascribing to the built-in electric field originated fromthe CuBi2O4/Mg-CuBi2O4 homojunction， opposite directionalmigration of both electrons on CB from pure CuBi2O4 to MgdopedCuBi2O4 and holes on VB from Mg-doped CuBi2O4 topure CuBi2O4 was realized. Meanwhile， the reduced AgP2W18during the reaction can store protons to satisfy the requirementof electrical neutrality， which then suppressed the interfacialcharge carrier recombination and facilitated thephotoelectrochemical reaction. Such synergetic effect betweenthe bulk and surface transport then not only promoted the chargecarrier separation and suppressed the recombination but alsoimproved the reaction kinetics and intensified the surfaceactivity. Thus， state-of-the-art photoelectrochemical activity hasbeen achieved. The combination of homojunction design andpolyoxometalate modification revealed an appealing avenue toprepare highly efficient CuBi2O4-based photocathode.

Author Contributions： Conceptualization， W.F. and H.F.;Methodology， W.F. and Y.Z.; Validation， W.F.， D.L. and H.F.;Formal Analysis， W.F.， D.L. and H.F.; Investigation， W.F.;Resources， D.L. and Q.L.; Data Curation， H.F. and D.L.;Writing-Original Draft Preparation， W.F. and H.F.; Writing-Review amp; Editing， H.F. and Q.L.; Supervision， Q.L.; ProjectAdministration， Q.L.; Funding Acquisition， Q.L.

Supporting Information： available free of charge via the internet at http：//www.whxb.pku.edu.cn.

References

（1） Rajeshwar， K.; Hossain， M. K.; Macaluso， R. T.; Janaky， C.; Varga，A.; Kulesza， P. J. J. Electrochem. Soc. 2018， 165， H3192.doi： 10.1149/2.0271804jes

（2） Hu， Y.; Wang， J.; Huang， H.; Feng， J.; Liu， W.; Guan， H.; Hao， L.;Li， Z.; Zou， Z. J. Mater. Chem. A 2023， 11， 149.doi： 10.1039/d2ta07117g

（3） Niu， J.; Peng， Y.; Hu， T.; Chen， Y.; Cheng， J.; Hu， Y. Chem. Eng. J.2023， 454， 140203. doi： 10.1016/j.cej.2022.140203

（4） Bruziquesi， C. G. O.; Stolzemburg， M. C. P.; de Souza， R. R.;Rodriguez， M.; Rocco， M. L.; Salom?o， P. E. A.; Nogueira， A. E.;López-Caba?a， Z. E.; Pereira， M. C.; Silva， A. C. Int. J. Hydrog.Energy 2023， 48， 3456. doi： 10.1016/j.ijhydene.2022.10.195

（5） Qu， L.; Tan， R.; Sivanantham， A.; Kang， M. J.; Jeong， Y. J.; Seo， D.H.; Kim， S.; Cho， I. S. J. Energy Chem. 2022， 71， 201.doi： 10.1016/j.jechem.2022.03.013

（6） Gu， M.; Yan， Z.; Wu， X.; Li， Z.; Dong， Y.; Wang， G. L. Nanoscale2023， 15， 2954. doi： 10.1039/d2nr05588k

（7） Bi， Y.; Tan， G. Q.; Zhang， B. X.; Yang， Q.; Feng， S. J.; Liu， C. J.;Liu， W. L.; Xia， A.; Ren， H. J.; Liu， Y. J. Alloy. Compd. 2023， 941，168992. doi： 10.1016/j.jallcom.2023.168992

（8） Sun， M.; Chen， W.; Jiang， X.; Liu， B.; Tan， B.; Luo， L.; Xie， M.;Zhang， Z. ACS Appl. Mater. Interfaces 2022， 14， 43946.doi： 10.1021/acsami.2c12309

（9） Gopannagari， M.; Reddy， D. A.; Hong， D. H.; Joshi Reddy， K. A.;Kumar， D. P.; Ahn， H. S.; Kim， T. K. J. Mater. Chem. A 2022， 10，6623. doi： 10.1039/d1ta09956f

（10） Bozheyev， F.; Ellmer， K. J. Mater. Chem. A 2022， 10， 9327.doi： 10.1039/d2ta01108e

（11） Lee， J.; Yoon， H.; Kim， S.; Seo， S.; Song， J.; Choi， B. U.; Choi， S. Y.;Park， H.; Ryu， S.; Oh， J.; et al. Chem. Commun. 2019， 55， 12447.doi： 10.1039/c9cc06092h

（12） Monny， S. A.; Zhang， L.; Wang， Z.; Luo， B.; Konarova， M.; Du， A.;Wang， L. J. Mater. Chem. A 2020， 8， 2498. doi： 10.1039/c9ta10975g

（13） Tan， B.; Reyes， A. M.; Menéndez-Proupin， E.; Reyes-Lillo， S. E.; Li，Y.; Zhang， Z. ACS Energy Lett. 2022， 7， 3492.doi： 10.1021/acsenergylett.2c01750

（14） Wang， X.; Liu， X.; Wu， Y.; Fu， Y.; Zhang， H.; Zhou， M.; Wang， Y.Appl. Catal. B-Environ. 2023， 323， 122182.doi： 10.1016/j.apcatb.2022.122182

（15） Lei， W.; Yu， Y.; Zhang， H.; Jia， Q.; Zhang， S. Mater. Today 2022， 52，133. doi： 10.1016/j.mattod.2021.10.028

（16） Fan， S.; Chen， Y.; Yang， L. J. Phys. Chem. C 2022， 126， 19446.doi： 10.1021/acs.jpcc.2c05593

（17） Sauty， M.; Lopes， N. M.; Banon， J.-P.; Lassailly， Y.; Martinelli， L.;Alhassan， A.; Chow， Y. C.; Nakamura， S.; Speck， J. S.; Weisbuch， C.Phys. Rev. Lett. 2022， 129， 216602.doi： 10.1103/PhysRevLett.129.216602

（18） Hasanvandian， F.; Zehtab Salmasi， M.; Moradi， M.; Farshineh Saei，S.; Kakavandi， B.; Rahman Setayesh， S. Chem. Eng. J. 2022， 444，136493. doi： 10.1016/j.cej.2022.136493

（19） Wang， Y.; Wang， H.; He， T. Chemosphere 2021， 264， 128508.doi： 10.1016/j.chemosphere.2020.128508

（20） Li， C.; He， J.; Xiao， Y.; Li， Y.; Delaunay， J.-J. Energy Environ. Sci.2020， 13， 3269. doi： 10.1039/d0ee02397c

（21） Kumar， M.; Meena， B.; Subramanyam， P.; Suryakala， D.;Subrahmanyam， C. NPG Asia Mater. 2022， 14， 88.doi： 10.1038/s41427-022-00436-x

（22） Hota， P.; Das， A.; Maiti， D. K. Int. J. Hydrog. Energy 2023， 48， 523.doi： 10.1016/j.ijhydene.2022.09.264

（23） Zang， D.; Wang， H. Polyoxometalates 2022， 1， 9140006.doi： 10.26599/POM.2022.9140006

（24） Gao， X.; Wang， J.; Xue， Q.; Ma， Y.-Y.; Gao， Y. ACS Appl. Nano Mater. 2021， 4， 2126. doi： 10.1021/acsanm.0c03406

（25） Ma， K.; Dong， Y.; Zhang， M.; Xu， C.; Ding， Y. J. Colloid Interface Sci. 2021， 587， 613. doi： 10.1016/j.jcis.2020.11.018

（26） Yu， L.; Liu， Q.; Ding， S.; Yu， J.; Peng， S.; Zhang， J.; Jiang， C.; Yang，G. Appl. Surf. Sci. 2022， 602， 154095.doi： 10.1016/j.apsusc.2022.154095

（27） Xin， J.; Pang， H.; Jin， Z.; Wu， Q.; Yu， X.; Ma， H.; Wang， X.; Tan， L.;Yang， G. Inorg. Chem. 2022， 61， 16055.doi： 10.1016/j.jallcom.2022.164907

（28） Du， J.; Ma， Y.-Y.; Cui， W.-J.; Zhang， S.-M.; Han， Z.-G.; Li， R.-H.;Han， X.-Q.; Guan， W.; Wang， Y.-H.; Li， Y.-Q. Appl. Catal.B-Environ. 2022， 318， 121812. doi： 10.1016/j.apcatb.2022.121812

（29） Li， N.; Liu， J.; Dong， B. X.; Lan， Y. Q. Angew. Chem. Int. Ed. 2020，59， 20779. doi： 10.1002/anie.202008054

（30） Wang， L.; Zhang， Q.; Wei， T.; Li， F.; Sun， Z.; Xu， L. J. Mater. Chem.A 2021， 9， 2912. doi： 10.1039/d0ta10303a

（31） Wang， L.; Li， M.; Zhang， Q.; Li， F.; Xu， L. Inorg. Chem. Front. 2021，8， 3566. doi： 10.1039/d1qi00503k

（32） Keshipour， S.; Asghari， A. Int. J. Hydrog. Energy 2022， 47， 12865.doi： 10.1016/j.ijhydene.2022.02.058

（33） Ogiwara， N.; Iwano， T.; Ito， T.; Uchida， S. Coord. Chem. Rev. 2022，462， 214524. doi： 10.1016/j.ccr.2022.214524

（34） McHugh， P. J.; Stergiou， A. D.; Symes， M. D. Adv. Energy Mater.2020， 10， 2002453. doi： 10.1002/aenm.202002453

（35） Zhang， M.; Li， H.; Zhang， J.; Lv， H.; Yang， G.-Y. Chin. J. Catal.2021， 42， 855. doi： 10.1016/S1872-2067（20）63714-7

（36） Contant， R.; Klemperer， W. G.; Yaghi， O. Inorg. Synth. 1990， 27，104. doi： 10.1002/9780470132586.ch18

（37） Ross-Medgaarden， E. I.; Wachs， I. E. J. Phys. Chem. C 2007， 111，15089. doi： 10.1021/jp074219c

（38） Xu， Y. X.; Jian， J.; Li， F.; Liu， W.; Jia， L. C.; Wang， H. Q. J. Mater.Chem. A 2019， 7， 21997. doi： 10.1039/c9ta07892d

（39） Sun， X.; Li， N.; Wang， X.; Mu， Y.; Su， C.; Cong， X.; Wang， X.; Wu，F(xiàn).; Wu， G.; Chen， X. Biomater. Adv. 2022， 143， 213179.doi： 10.1016/j.bioadv.2022.213179

（40） Kumar， M.; Ghosh， C. C.; Meena， B.; Ma， T. Y.; Subrahmanyam， C.Sustain. Energy Fuels 2022， 6， 3961. doi： 10.1039/d2se00600f

（41） Li， T.; Li， M.; Jiang， J.; Zhao， Z.; Li， Z.; Zhao， C.; Wang， X.; Dong，S. Appl. Catal. B-Environ. 2023， 122539.doi： 10.1016/j.apcatb.2023.122539

（42） Liu， S.; Zhu， L.; Cao， W.; Li， P.; Zhan， Z.; Chen， Z.; Yuan， X.; Wang，J. J. Alloy. Compd. 2021， 858， 157654.doi： 10.1016/j.jallcom.2020.157654

（43） Cao， J.; Cen， W.; Jing， Y.; Du， Z.; Chu， W.; Li， J. Chem. Eng. J.2022， 435， 134683. doi： 10.1016/j.cej.2022.134683

（44） Zhang， X.; Dai， J.; Ding， J.; Tan， K. B.; Zhan， G.; Huang， J.; Li， Q.Catal. Sci. Technol. 2022， 12， 2426. doi： 10.1039/d1cy02185k

（45） Wu， Z.; Zhang， Z.; Sun， M.; Tan， B.; Liu， B.; Han， W.; Xie， E.; Li， Y.Adv. Mater. Interfaces 2021， 8， 2101443.doi： 10.1002/admi.202101443

（46） Xu， N.; Li， F.; Gao， L.; Hu， H.; Hu， Y.; Long， X.; Ma， J.; Jin， J. ACSSustain. Chem. Eng. 2018， 6， 7257.doi： 10.1021/acssuschemeng.7b04133

（47） Xu， X.; Li， Y.; Liu， C.; Zhang， P.; Fan， K.; Wu， X.; Shan， Y.; Li， F.Dalton Trans. 2023， 52， 5848. doi： 10.1039/d3dt00542a

（48） Chen， Z.; Dinh， H. N.; Miller， E. Photoelectrochemical WaterSplitting Standards， Experimental Methods， and Protocols; Springer：New York， NY， USA， 2013; pp. 63–68.

（49） Hu， Y.; Hu， Q.; Qi， Y.; Zhang， W.; Liu， C.; Wang， Y.; Guan， H.; Hao，L. Dalton Trans. 2023， doi： 10.1039/d2dt04009c

（50） Yuan， N.; Zhang， J.; Zhang， S.; Chen， G.; Meng， S.; Fan， Y.; Zheng，X.; Chen， S. J. Phys. Chem. C 2020， 124， 8561.doi： 10.1021/acs.jpcc.0c00422

（51） Gu， L.; Hou， X.; Lei， Y.; Gou， S.; Yang， X.; He， W.; Zheng， Z.J. Alloy. Compd. 2022， 904， 163934.doi： 10.1016/j.jallcom.2022.163934

（52） Wei， S.; Wang， C.; Long， X.; Wang， T.; Wang， P.; Zhang， M.; Li， S.;Ma， J.; Jin， J.; Wu， L. Nanoscale 2020， 12， 15193.doi： 10.1039/d0nr04473c

（53） Lamers， M.; Sahre， M.; Müller， M. J.; Abou-Ras， D.; van de Krol， R.;Abdi， F. F. APL Mater. 2020， 8， 061101. doi： 10.1063/5.0003005

（54） Berglund， S. P.; Abdi， F. F.; Bogdanoff， P.; Chemseddine， A.;Friedrich， D.; van de Krol， R. Chem. Mater. 2016， 28， 4231.doi： 10.1021/acs.chemmater.6b00830

（55） Wang， J.; Ni， G.; Liao， W.; Liu， K.; Chen， J.; Liu， F.; Zhang， Z.; Jia，M.; Li， J.; Fu， J. Angew. Chem. Int. Ed. 2022， 62， e202217026.doi： 10.1002/anie.202217026

（56） Wang， J.; Zhao， C.; Yuan， S.; Li， X.; Zhang， J.; Hu， X.; Lin， H.; Wu，Y.; He， Y. J. Colloid Interface Sci. 2023， 638， 427.doi： 10.1016/j.jcis.2023.02.005

（57） Pakdel， A.; Khan， A. U.; Pawula， F.; Hébert， S.; Mori， T. Adv. Mater.Interfaces 2022， 9， 2200785. doi： 10.1002/admi.202200785

（58） Li， M.; Tian， X.; Zou， X.; Han， X.; Du， C.; Shan， B. Int. J. Hydrog.Energy 2020， 45， 15121. doi： 10.1016/j.ijhydene.2020.03.242

（59） Fang， W.; Tao， R.; Jin， Z.; Sun， Z.; Li， F.; Xu， L. J. Alloy. Compd.2019， 797， 140. doi： 10.1016/j.jallcom.2019.05.063

（60） Baerends， E. J. Phys. Chem. Chem. Phys. 2017， 19， 15639.doi： 10.1039/c7cp02123b

（61） Dong， W. J.; Navid， I. A.; Xiao， Y.; Lee， T. H.; Lim， J. W.; Lee， D.;Jang， H. W.; Lee， J. L.; Mi， Z. J. Mater. Chem. A 2022， 10， 7869.doi： 10.1039/d2ta00032f

（62） Tan， B.; Liu， B.; Sun， M.; Li， Y.; Cao， Z.; Zhang， Z. J. Mater. Chem.A 2022， 10， 9427. doi： 10.1039/D2TA00476C

（63） Yang， L.; Wang， J.; Ma， T.; Zhang， L. J. Colloid Interface Sci. 2022，611， 760. doi： 10.1016/j.jcis.2021.11.100

（64） Jiang， W.; Zhang， L.; Ni， C.; Shi， M.; Zhao， Y.; Deng， Y.; Chi， H.;Chen， R.; Wang， X.; Li， R.; et al. Angew. Chem. Int. Ed. 2023， 62，e202302575. doi： 10.1002/anie.202302575

（65） He， H.; Huang， X.; Liu， C.; Li， D.; Chen， S.; Yan， Z.; Liu， Y. Appl.Surfaces Sci. 2023， 610， 155493. doi： 10.1016/j.apsusc.2022.155493

（66） Mary， A. S.; Murugan， C.; Pandikumar， A. J. Colloid Interface Sci.2022， 608， 2482. doi： 10.1016/j.jcis.2021.10.172

（67） Chen， J. J.; Vila-Nadal， L.; Sole-Daura， A.; Chisholm， G.; Minato， T.;Busche， C.; Zhao， T.; Kandasamy， B.; Ganin， A. Y.; Smith， R. M.; etal. J. Am. Chem. Soc. 2022， 144， 8951. doi： 10.1021/jacs.1c10584

（68） Gao， L.; Li， F.; Hu， H.; Long， X.; Xu， N.; Hu， Y.; Wei， S.; Wang， C.;Ma， J.; Jin， J. ChemSusChem 2018， 11， 2502.doi： 10.1002/cssc.201800999

（69） Klahr， B.; Gimenez， S.; Fabregat-Santiago， F.; Hamann， T.; Bisquert，J. J. Am. Chem. Soc. 2012， 134， 4294. doi： 10.1021/ja210755h

國(guó)家重點(diǎn)研發(fā)計(jì)劃（2021YFF0500700）， 國(guó)家自然科學(xué)基金（51976090， 52006101， 52006103）， 江蘇省碳達(dá)峰碳中和科技創(chuàng)新項(xiàng)目（BE2022024）及江蘇省自然科學(xué)基金（BK20200491）資助